summaryrefslogtreecommitdiff
path: root/lib/csu
AgeCommit message (Collapse)Author
2003-04-17i386 startup code for ELF. based on newer version of NetBSD code.Dale Rahn
2003-03-31add a newline at end of file. gcc3 rightfully complains if the fileBruno Rohee
is not newline terminated. From Thorsten Glaser
2003-03-10duplicate words and spelling fixes in commentsDavid Krause
ok miod@
2003-02-28copy basename of argv[0] to bss for __progname, so that large stack smashTheo de Raadt
does not make propolice reporting ineffective; millert miod ok
2003-02-26Switch to a C version of powerpc crt0 file. easier to read. movesDale Rahn
_progname storage into data instead of stack.
2003-01-16extern the eprol and etext; better eprol definitionMichael Shalayeff
2002-11-10build crtbegin/crtend as it is ment for elfMichael Shalayeff
2002-09-17uncommit, since it breaks macppc libc. millert and miod have now said theyTheo de Raadt
did NOT approve those for commit. why did mickey feel he was ok to go commiting a set of diffs which had not been passed around and tested by people? we don't know. mickey, have you got something to say for yourself?
2002-09-17use a different name for __syscall(); millert@ miod@ niklas@ okMichael Shalayeff
2002-08-11Adapt sparc crt* to how things are done on our other ELF architectures.Artur Grabowski
Some parts of crt0.c are from NetBSD.
2002-07-22Remove some dead ifdefs to make the code more readable.Artur Grabowski
2002-07-14Fix an oversight in revision 1.9. ld_entry was set in one more spot.Artur Grabowski
2002-07-11We need to map the data segment of ld.so with PROT_EXEC for the GOT.Artur Grabowski
This is in preparation for actually honoring exec protection in sparc pmap. NOTICE! You really want to rebuild the whole userland with this change before the kernel fix goes in. Otherwise all shared binaries will dump core.
2002-06-05one more level of indirection for dlopen structure.Marc Espie
Allows dlopen to be used much earlier, e.g., in dynamic constructors. From NetBSD. okay art@
2002-02-17Manual cleanup of remaining userland __P use (excluding packages maintained ↵Todd C. Miller
outside the tree)
2002-02-16Part one of userland __P removal. Done with a simple regexp with some minor ↵Todd C. Miller
hand editing to make comments line up correctly. Another pass is forthcoming that handles the cases that could not be done automatically.
2002-02-02fix some protosMichael Shalayeff
2001-09-11add sparc64; op2@squish.orgJason Wright
2001-09-02I don't even want to know how I could screw this up so bad and noone noticed.Artur Grabowski
2001-08-29build crt0, crtbegin and friends.Artur Grabowski
2001-08-29crt0. Almost like on alpha, except that sparc64 needs an assembler stubArtur Grabowski
for the moment. (why make the ABI so complicated?) The assembler is inspired by crt0 in NetBSD.
2001-08-19remove unused bits and pieces.Brad Smith
-- Ok'd by art@
2001-08-19make obj in hppa alwaysMichael Shalayeff
2001-08-19either way this is what i'm using nowMichael Shalayeff
2001-08-12#(endif|else) foo is incorrect, make it #endif /* foo */Heikki Korpela
deraadt@ ok
2001-07-18WallTheo de Raadt
2001-07-11remove mips leftovers; ok bradPeter Valchev
2001-06-26Put m88k-related files in the correctly named directory.Miod Vallat
2001-06-01proto for dlerror() changedTheo de Raadt
2001-05-29fix the alpha after that last diff which was obviously not testedTheo de Raadt
2001-05-28Commonize csu code for elf systems, powerpc now no longer has it's ownDale Rahn
versions of these files. Fixed a bug in ld.so in this, instead of scheduling the fini of each of the shared libraries with atexit. schedule a function of ld.so itself and it will walk all of the open libraries when the program exits. otherwise a shared library could be dl_open()ed and then dl_close()d and then it would not be mapped for the atexit processing. TODO: What if atexit is not found (process did not link against libc?) Do shared libraries that are dl_closed have their global destructors run?
2001-05-11MAP_COPY -> MAP_PRIVATEArtur Grabowski
2001-03-24Use "realinstall" to allow the framework to do other work if needed;Thorsten Lockert
path of least surprise. Ok millert@
2001-02-03Remove even more dead code.Artur Grabowski
2001-02-03oops.Artur Grabowski
2001-02-03Schedule running of __fini in __init, not __start.Artur Grabowski
2001-02-03An OpenBSD tag doesn't hurt.Artur Grabowski
2001-02-03Add an OS note identifying OpenBSD binaries.Artur Grabowski
This appears to be the standard way to do it.
2001-02-03Decruftification.Artur Grabowski
2001-02-03Use crtbegin and crtend from common_elf.Artur Grabowski
2001-02-03Create a common_elf directory with crtbegin and crtend that canArtur Grabowski
be shared between archs.
2001-02-03Don't define ECOFF_COMPAT. We're ELF.Artur Grabowski
2001-01-31deconfuse include path. (XXX - this for my future hacks).Artur Grabowski
2001-01-25This is not how we'll do dynamic libraries on alpha anyway.Artur Grabowski
Zap dead code.
2000-10-13Changes to the powerpc startup files to work in a more typical ELFDale Rahn
manner. the linker knows about the magic function _init and _fini and will create a .dynamic tag to point to those symbols in the executable/shared libraries. Take advantage of that fact to allow ld.so to easily find the appropriate ctors/dtors calls for shared libraries Since the crtbeginS.c now contains executable code, it must be compiled with -fPIC. Rename the __init and __fini functions to _init/_fini so the linker will locate them.
2000-06-28Powerpc was missing the scrt.o startup file. now used by the ramdisk instbin.Dale S. Rahn
also scrt does not support c++ ctors/dtors thus does not require crtbegin.o/crtend.o which it normally does.
2000-06-13Support files for shared library, expected by gcc, allows shared libraries toDale S. Rahn
have constructors and destructors.
1999-12-09dlopen and dlsym take const char * parameters.Marc Espie
Fix prototypes, propagate const where applicable.
1999-08-24Do not error out if we have been statically linked but would like ld.so for ↵Niklas Hallqvist
dlopen and it is not there.
1999-08-20Fix RCS IdsNiklas Hallqvist